Here is a project in which we use different Machine Learning and Deep Learning Algorithms to benchmark FashionMNIST Dataset and compare the results with original MNIST Dataset.
Classifier | Optimizer | Preprocessing | Fashion test accuracy | MNIST test accuracy | Location |
---|---|---|---|---|---|
2 Conv Layers with max pooling + 2 Connected Layer (PyTorch) | SGD + Nesterov | None | 92.69% | 99.34% | CNN-FashionMNIST Model A |
2 Conv Layers with max pooling + 1 Connected Layer (PyTorch) | SGD + Nesterov | None | 91.21% | 99.23% | CNN-FashionMNIST Model B |
3 Conv Layers with max pooling + 2 Connected Layer (PyTorch) | SGD + Nesterov | None | 93.18% | 99.37% | CNN-FashionMNIST Model C |
More Algorithms will be added soon.